Shuai Wei

Lead in-silico monoclonal and human-like single-domain antibody library design by protein language models.

Developed and implemented antibody biophysical/developability profilers based on DL transfer learning models that achieved state-of-the-art performance.

Developing and implementing multi-property co-optimization for antibody discovery using multi-task learning and Bayesian optimization.

Developed and implemented deep learning (DL)/machine learning (ML) model for de novo large molecule design.

Architected and implemented data generation pipeline from computational chemistry software.

Performed molecular dynamic simulation methods for evaluating and predicting large molecule developability.

Provided consultation for ML and DL research projects including natural language processing (NLP), computer vision (CV), hierarchical clustering using Tensorflow, Pytorch and Keras.

Taught tutorials and classes on machine learning, data analysis/visualization and version control tools.

Developed machine learning models to facilitate fast simulating protein folding and dynamics in complex environments.

Developed and implemented GPU functionalities using CUDA and OpenCL that speed up around 20x for a computational chemistry software CHARMM.

Renovated data analysis tools for complex biomolecule from nonlinear laser spectroscopy measurement.